Homologous recombinational repair (HRR) is essential for the accurate repair of inter-strand crosslinks (ICLs) and double stranded DNA breaks (DSBs) that can form from exposure to environmental toxins, ionizing radiation or endogenous metabolism. Essential to this process in humans are five RAD51 paralogs (RAD51B, RAD51C, RAD51D, XRCC2 and XRCC3), which carry out their roles through interaction with one another, RAD51 and DNA. However, the biological function of RAD51 paralog complexes are poorly understood, and despite years of study mechanistic information about their assemblies and activities has been enigmatic. Based on breakthrough preliminary data that includes robust soluble expression of RAD51 paralog complexes and a RAD51C homolog crystal structure, I propose two aims that combine small-angle X-ray scattering and macromolecular crystallography to build accurate RAD51 paralog subcomplexes. These structural models will be experimentally validated using yeast two-hybrid and protein pull-down assays. Results from this proposal, including robust methods for the production of soluble RAD51 paralogs and a molecular basis from which to design separation-of-function mutations, will pave the way for new biochemical and biological assays by myself and others in the field to dissect the function of RAD51 paralogs in HRR and other genome maintenance pathways. This structural work will also provide a molecular framework from which to understand cancer predisposing RAD51 paralog mutations that are emerging from the clinic. Furthermore, due to the sensitivity of RAD51 paralog-deficient cells to ionizing radiation and ICL agents, this work may provide a molecular framework to design specific RAD51 paralog inhibitors that could be useful for advanced therapies that work by synthetic lethality in combination with targeted cancer therapeutics. Collectively, results from this proposal will provide important reagents and informative structures that will lead to greater understanding of RAD51 paralog cellular functions, and provide a framework to understand and classify RAD51 paralog mutations from the clinic.